Download Conclusion: Reconstructing and Modelling Past Oceans 1 A

1
Conclusion: Reconstructing and Modelling Past Oceans
Paleoceanography of the Late Cenozoic (Elsevier)
Volume 1: Methods in Late Cenozoic Paleoceanography
Katrin J. Meissner
The science of reconstructing past climates has evolved rapidly over the past three
decades. Having started as a subdiscipline of geology and geochemistry [Imbrie and Imbrie, 1979], it is today an interdisciplinary research field which unifies a large international
scientific community. Some of the most innovative and suprising research in this field has
important implications for future climate change. As such, paleoscience is today recognized
as being of great relevance to societal concerns. For example, ice core analyzes revealed
that greenhouse gas levels are higher today than they have been for hundreds of thousands
of years. In other words, anthropogenic impacts are pushing the climate system towards a
state for which there is no “reference climate” in paleo records of the Quaternary. Therefore
a better understanding of the climate system and feedbacks within the climate system is
crucial for our future. Another disconcerting finding of paleoresearch is the fact that the
climate system is highly non-linear: in the past, rapid and large amplitude climate change
has occured in response to slowly varying, small amplitude forcing. It is critical that policy
makers understand what paleoscience has to say about rapid climate change in order to fully
understand the significance of future rapid climate change. The UN-sanctioned Intergovernmental Panel on Climate Change (IPCC) recognizes the importance of understanding past
climate change to predict future climate scenarios and has devoted an entire chapter in the
fourth assessment report (AR4) to paleoclimates.
The ocean’s high heat capacity and its ability to transport energy and to sequester and
release greenhouse gases give it an important role in helping to determine the state of the
planet’s climate. Compared to the atmosphere, the ocean has a long response time to
perturbations and this makes it a key player in climate change on timescales of hundreds
to ten thousands of years. Thus, paleoceanography is a crucial component of paleoclimatic
research.
1
A brief historical overview
The reconstruction of paleocean characteristics and dynamics requires climatic detective
work. It involves the dating and interpretation of paleoclimatic records as well as the definition of physical and dynamical constraints which specify possible circulation patterns and
characteristics. As a single drill core, analysis method or numerical model cannot possibly
address fundamental scientific questions regarding past climate change, international cooperative efforts are necessary to construct multiproxy records and evaluate climate models.
2
The first interdisciplinary project to reconstruct climate began in the 1970s and was led
by the marine geologists John Imbrie and Jim Hayes and the geochemist Nick Shackelton.
The CLIMAP (Climate: Long-range Investigation, Mapping, and Prediction) project reconstructed the surface of the Earth at the Last Glacial Maximum (LGM) at 21,000 years BP
[CLIMAP Project Members, 1976, 1981]. Although the reconstruction was based on ocean
sediment proxy data, the expertise of scientists from other disciplines was also sought-after
including researchers from the still very new field of climate modelling, for example.
The first climate models evolved from numerical weather prediction models and were
developed during the early 1960s. These atmosphere general circulation models (AGCMs)
had become the central tool of climate science by the 1970s. The initial focus of modeling
past climates was the Last Glacial Maximum (since it represented a very different climate
from today and there was a relatively rich amount of proxy data) at 21,000 BP. Some of the
first studies were simulations by Williams et al. [1974]; Gates [1976a, b] and Hansen et al.
[1984], amongst others. As atmosphere-only models, these simulations relied on boundary
conditions (CLIMAP sea surface temperatures, for example). An exhaustive overview of the
results of these early model studies can be found in Crowley and North [1991].
Meanwhile, ocean modellers were beginning to build similar models of the ocean (ocean
general circulation models or OGCMs). As soon as the first ocean models became operational, it became apparent that the poleward ocean heat transport plays a key factor in
climate change and variability [Manabe and Bryan, 1985; Bryan, 1986]. The studies of Bryan
[1986] and Manabe and Stouffer [1988] should be highlighted in this context as these early
OGCM simulations showed for the first time that multiple equilibria of the thermohaline
circulation could exist in more explicit models under identical boundary conditions. This
finding resulted in an exhaustive and still ongoing study of the hysteresis behaviour of the
thermohaline circulation in response to buoyancy forcing (see Weaver and Hughes [1992]
for a thorough early overview paper or some of the more recent studies (Rahmstorf [1995];
Weaver [1995]; Ganopolski and Rahmstorf [2001]; Rahmstorf [2002]; Rahmstorf et al. [2005]
amongst many others).
In the late 1980s, the COHMAP (Cooperative Holocene Mapping Project) project took
the place of CLIMAP in again successfully combining skills of many disciplines into a single
interdisciplinary effort. The objective of the COHMAP project was to reconstruct climate
conditions at successive 3,000 years intervals from 18,000 BP to present. As climate models
had evolved drastically in the 1980s, this time a variety of proxy data sources was used in combination with model simulations to reconstruct paleo environments. Therefore COHMAP
was the first interdisciplinary effort to promote the interaction between climate modellers
and Quaternary Scientists. Broad data sets were used to define model boundary conditions,
in order to test model accuracy and improve our understanding of external and internal
climate forcing mechanisms [COHMAP Members, 1988].
The next deciding step forward in the history of climate modelling of past environments
was the consideration of biogeochemical cycles as a crucial component of the climate system. Until the early 1990s, the main focus of climate modelling was on physical feedbacks
between climate components (mainly the ocean, atmosphere and sea ice), but a few researchers had begun to investigate the importance of biogeochemical cycles as early as the
mid-1980s. These first simulations were carried out using 3-box models [Knox and McElroy,
3
1984; Sarmiento and Toggweiler, 1984; Siegenthaler and Wenk, 1984] and highlighted an
important aspect of the marine carbon cycle: The CO2 content of the warm (low latitude)
surface ocean and hence also of the atmosphere is regulated by the surface waters of the cold
high-latitude box. This finding is referred to as the Harvardton-Bear effect after the authors’
three institutions (Havard, Princeton and Bern) and has important implications for paleoclimate conditions with reduced atmospheric CO2 [Broecker et al., 1999]. The growing evidence
over the last 20 years is that a better quantification of biogeochemical interactions is critical
in the understanding of climate change and variability and this has led to the development
of biogeochemical processes in box models [Stephens and Keeling, 2000; Toggweiler, 1999;
Gildor et al., 2002] and OGCMs. As a result, the first paleoclimate model studies including
oceanic and/or global biogeochemical processes in OGCMs started to appear in the early
1990s [Heinze et al., 1991] and are still being developed (Schmittner [2005] amongst others).
Today there are several interdisciplinary alliances which aim to catalyze and compare
studies from a wide range of different research groups with a diverse set of research tools.
These groups include the Coupled Carbon Cycle Climate Model Intercomparison Project
(C4MIP), PAGES (Past Global Changes), RAPID (Rapid Climate Change) and PMIP (Paleoclimate Modeling INtercomparison Project) which will be discussed in Section 4.
2
Classification of Climate Models
The theoretical approach to paleoceanography uses quantitative ocean and climate models
to reconstruct paleoconditions of the ocean and interpret observations. A number of different
ocean and coupled climate models have been used in the past and a brief overview of the
hierarchy and use of these models in the broad research area of paleoceanography will now
be given (for an extensive overview of numerical climate models, the reader is referred to
McGuffie and Henderson-Sellers [2001]).
2.1
Conceptual Models
Climate Models can be split into two categories, conceptional and numerical models. Whereas
conceptual models provide non- or semi-quantitive, nondynamic simulations of paleoclimate,
numerical models are quantitative and can be dynamic [Parrish, 1998]. One of the first paleoclimate studies using a conceptual model is described in Nairn and Smithwick [1976]. Other
purely conceptual models of the atmosphere [Parrish, 1982] and ocean [Ziegler et al., 1981]
were used for paleoclimate studies in the late 1970s and early 1980s. The next step up in
the hierarchy of climate models are the numerical conceptual models. These models are still
very simple and take only a few processes into account, but make use of computer power.
To understand the usefulness of these simple models, consider the following example.
The 100,000 year rhythm of the major glacial-interglacial cycles during the past 700,000
years is thought to be tied to the changing eccentricity of the Earth’s orbit (which has
characteristic periods of 95,000, 131,000, 413,000 and 2.1 million years) in what is commonly
referred to as Milankovitch theory. Challengers of this theory argue that these changes
result in very small changes in insolation and are unlikely to drive the climate system. Also,
4
the 413,000 period of eccentricity resulted in minimal insolation changes during the warm
interglacial episode about 400,000 years ago. The proxy records show an especially warm
interglacial during that time contrary to all predictions solely based on insolation in the
Northern Hemisphere. Paillard [1998] developed two simple multi-equilibria ice models which
were able to give more insight into Milankovitch forcing and the warm interglacial episode.
With his two very simple multi-equilibria models, Paillard [1998] successfully simulated both
- the transition from a 41 to 100 kyr period in the glacial-interglacial cycle as well as a warm
interglacial 400,000 years ago.
2.2
Numerical Models
It is, of course, impossible to build an exact model of the climate system; for the ocean alone,
the position and momentum of approximately 5x1046 molecules would have to be calculated
at each instant of time. Instead, the ocean, atmosphere, sea ice, or land surface is split into
discrete macroscopic elements with measurable characteristics such as temperature, density,
velocity, etc. The state of and exchange between these discrete elements follow physical laws
and can therefore be determined with numerical models. Small scale processes within each
element can also influence the large scale pattern and therefore have to be parametrized.
Existing ocean (and climate) models differ in regards to:
• their temporal and spatial resolution (resolution is defined as the spatial scale which
defines the boundary between processes that are resolved by the model and those which
are parametrized)
• the nature of processes which are resolved (for example, some models include biogeochemical cycles whereas other models resolve physical processes only)
• the number of subsystems taken into account (for example, an ocean model needs
boundary conditions at the surface which can either be provided by data, or by an
atmosphere model which is physically interacting with the ocean model)
As the different subsystems (ocean, atmosphere, sea ice, continental ice sheets, vegetation,
etc.) of the climate system interact with each other in a complex manner and on a very
broad range of timescales, climate modelling reduces to the process of identifying isolable
subsystems and processes that are relevant to the problem at hand. While identifying these
subsystems and processes, the researcher has also to keep in mind that these processes have
to be suitable to be simulated by limited mathematical models and will provide results in a
reasonable computational time [Crowley and North, 1991].
The simplest class of climate models includes one-dimensional Energy Balance Models
which were first developed by Budyko [1969] and Sellers [1969]. It is interesting to note that
these simple models yield two stable solutions under present day boundary conditions: the
present day climate and a completely frozen Earth (also called “Snowball Earth”).
2.2.1
General Circulation Models
Because the boundary data for paleoclimatic simulations tend to contain large uncertainties,
global models are better suited for this research area than regional models. The most com-
5
prehensive results are of course given by global general circulation models (GCMs). GCMs
consist of a three-dimensional representation of the ocean (or atmosphere) and are to date
the most complex models available. They are built on the conservation equations of momentum, mass and tracers and can include parametrizations of the carbon cycle or atmospheric
chemistry. When run in an uncoupled mode (atmosphere or ocean model only) they rely on
reconstructed data specifying the boundary conditions at the bottom of the atmosphere (or
surface of the ocean). For example, as already mentioned in Section 1, numerous modelling
studies restored the ocean surface characteristics to the CLIMAP data set [CLIMAP Project
Members, 1976] for simulations of the Last Glacial Maximum (e.g. Fichefet et al. [1994]).
A better approach is to use coupled ocean-atmosphere models; by computing the surface
boundary conditions, one can bypass the data problem. For example, atmosphere general
circulation models (AGCMs) coupled to a mixed-layer ocean are still commonly used. This
type of model predicts changes in sea-surface temperatures and sea ice by treating the ocean
as though it were a layer of water of constant depth (typically 50 meters), heat transports
within the ocean being specified and remaining constant while climate changes. With a constant heat transport in the ocean, these models are quite limited for simulations under very
different climatic conditions [Bice et al., 2000]. The most complex models in use today are
coupled atmosphere-ocean general circulation models (AOGCMs, e.g. Bush and Philander
[1998]). Some recent models include the biosphere, carbon cycle and atmospheric chemistry as well. However, coupled ocean-atmosphere GCMs often need flux adjustments. Flux
adjustments balance surface fluxes at the ocean-atmosphere interface to avoid a numerical
drift of the coupled system. As flux adjustments have been “tuned” to the present day climate, the use of these adjustments to simulate past climates is not very reliable. However,
some recent studies use coupled atmosphere-ocean GCMs which do not need artificial flux
adjustments (e.g. LeGrande et al. [2006]).
2.2.2
Earth System Models of Intermediate Complexity
Earth System Models of Intermediate Complexity (EMICs, Claussen et al. [2002]) have been
developed recently to bridge the gap between simple, conceptual models (such as box models
or EBMs) and comprehensive general circulation models (GCMs). EMICs include most of
the processes described in comprehensive models, but in a more reduced (parametrized) form.
As climate subsystems other than the atmosphere and ocean also play an important role in
climate change (continental ice sheets, sea ice and land surface processes, for example), the
class of Earth System Models explicitly simulates the interactions among several components
of the natural Earth system. The advantage of EMICs lies in their computational efficiency,
which allows for long-term climate simulations over several tens to thousands of years. Also,
the large number of processes described in EMICs enables the user to investigate interactions
and feedbacks within the climate system in a broad range of sensitivity experiments [Kubatzki
and Claussen, 2003]. EMICs are used to explore the parameter space, develop analysis
methods, and analyze the response of climate over long timescales. The modelling community
is continuously integrating new processes and subsystems in their models to obtain a better
representation of the climate system dynamics. For example, there is growing evidence
that biogeochemical interactions between the subsystems are more important than initially
6
thought (carbon cycle, nitrogen cycle, methane, etc.) and most of the EMICs used as of today
include some parametrization of oceanic and even global biogeochemical cycles (Bern 2.5D
climate model, CLIMBER-2, CLIMBER 3α, GENIE, ISAM-2, LOVECLIM, MIT Integrated
Global System Model, MoBidiC, UVic Earth System Model).
Earth System Models of Intermediate Complexity are today widely used for paleoceanographic and paleoclimate simulations.
3
Models and Proxy Data
The interpretation of paleoproxy data is an ongoing challenge for paleoclimate scientists. As
a striking example, one could compare the studies of Clark et al. [2002] and Bond et al. [2001].
Both papers are highly regarded, yet draw opposite conclusions from the atmospheric δ 14 C
record. Whereas Bond et al. [2001] relates the variability of atmospheric δ 14 C to changes
in solar radiation, Clark et al. [2002] interprets the same type of record as a signature of
variability in the thermohaline circulation and ocean heat transport. On the other hand,
model simulations of past climates depend strongly on boundary conditions, assumptions
and the model used for the study. The simulated climate for a certain time span can be
radically different depending on the model and boundary conditions used. Interpretation of
measured paleoclimate data is thus urgently needed through collaboration between modellers
and observers.
3.1
Data Assimilation
One possible approach to combine proxy data with climate models is that of “data assimilation”. Data assimilation involves the construction of a field that accommodates best the
information obtained from paleoproxies with the physical (and dynamical) constraints of the
climate system using coupled climate models (e.g. Paul and Schäfer-Neth [2005]). Thus
the models help to physically interpret the rather local proxy data in a global context. Inverse models, for example, seek for a compromise between the hydrographic and paleoproxy
data. They determine circulation and biogeochemical characteristics by fitting the model to
datasets of temperature, salinity, oxygen, nutrients, radiocarbon etc. For example, Winguth
et al. [2002] used an inverse method [Wunsch, 1997] by which an OGCM (the Hamburg
large-scale geostrophic OGCM LSG) can be adjusted to fit the paleonutrient data in order
to reproduce ocean circulation and surface boundary conditions for the Last Glacial Maximum (LGM). The results of this assimilation study confirm that the North Atlantic Deep
Water was shallower and the North Atlantic Intermediate Water stronger during the LGM
compared to present day. The optimized glacial circulation pattern in the Southern Ocean
are consistent with the Cd/Ca measurements but cannot explain changes in δ 13 C.
Other modellers use a more empirical approach by realizing several model simulations
and comparing the simulated characteristics to paleodata. For example, Schmittner et al.
[2002] compared simulated changes in sea surface properties for a large variety of overturning strengths with different reconstruction sets. In this study, the results were found to depend strongly on the data set used. Sea surface temperature reconstructions from CLIMAP
7
[CLIMAP Project Members, 1976, 1981] and earlier salinity reconstructions based on planktonic foraminifera [Seidov et al., 1996; Duplessy et al., 1991; Sarnthein et al., 1995] are most
consistent with a significant reduction of the circulation, while recent reconstructions using
dynocyst assemblages [de Vernal et al., 2000] allow no unequivocal conclusion.
The disconcerting results of these two examples point to one of the most important challenges one encounters when combining paleoproxy data and climate models: the scale difference. Climate models reproduce large-scale features (reflecting global differential heating,
the impact of large-scale topography, etc.) whereas smallscale features (secondary mountain ranges, marginal seas, etc.) are not well described because of insufficient resolution
[Renssen et al., 2004]. On the other hand, the spatial and temporal sampling of proxy data
is sparse. Therefore, the local character of proxy records might show the signature of local
events which are not resolved in a global model with coarse resolution. Also, the proxy data
usually reflects seasonal phenomena which might be mis-interpretated as representing annual
averages.
The uncertainties of paleoproxy reconstructions combined with the highly non-linear
nature of the climate system makes a complete agreement between models and proxy data
impossible. Therefore, palaeodata assimilation schemes may be seen as an “estimate of the
probability distribution functions of past climatic features and model parameters” [Crucifix,
2005b].
3.2
Integrating proxy data into models
Another possible approach is to incorporate paleoproxy data (δ 18 O, δD, δ 14 C, δ 13 C, δ 10 Be,
etc.) as prognostic active tracers in climate models. These climate models can then be run
under constant forcing to reproduce pattern of proxy data for a certain equilibrium climate
state (e.g. the Last Glacial Maximum) which can be compared to observed proxy data
obtained from ice cores, marine sediments and other records. Additionally, time-dependent
(or transient) experiments can be performed under time-varying forcings (such as meltwater
events or changing solar activity) to produce timeseries which can be directly compared to the
proxy record. This has been done with uncoupled ocean (or atmosphere) general circulation
models (e.g. Schmidt [1999]; Werner et al. [2000]; Butzin et al. [2005]), vegetation models
[Kaplan et al., 2002] and continental ice sheet models (e.g. Clarke et al. [2005]). However,
the importance of interactions between atmosphere, oceans and other systems such as the
biosphere and the cryosphere point to the necessity of using coupled models. To date,
there have been only a few studies simulating paleoproxy data with either coupled oceanatmosphere GCMs or Earth System Models (e.g. LeGrande et al. [2006]; Crucifix [2005a];
Roche et al. [2004]; Meissner et al. [2003]; Stocker and Wright [1996]). These quantitative
comparisons between climate models and paleoproxies are one of the key foci in paleoclimate
modelling as they will allow for improved interpretation of the paleoproxy records.
A third way to bridge the gap between the modelling and proxy data approach is to
find locations of proxy data records of special interest with the use of climate models. A
simulation including prognostic paleoproxy tracers can determine the geographical region of
greatest impact on a given paleoproxy data during a given climate event (e.g. Meissner et al.
[2003]).
8
3.3
Testing and evaluating models
Finally, geological data provide us with the only way of verifying climate models. Even
though paleoproxy data are point measurements whereas climate models operate on large
grid cells, paleoproxy time series offer anchor points to test and evaluate models. In addition, sensitivity experiments can be performed to describe mechanisms which assess the
model response. Thus, numerical simulations of past climates constitute an independent
test of the performance and reliability of the climate models that are used for future climate
change scenarios [Crucifix et al., 2005].
In conclusion, large amounts of paleoproxy data have been retrieved from various types
of archives, but attempts to use numerical models for verification and interpretation of this
data are sparse. The science of using three-dimensional climate models to interpret paleo
records is still in its infancy.
4
International Programs
Today, there are several international and interdisciplinary programs to coordinate and encourage the systematic study of past climates. For example, the Paleoclimate Modeling Intercomparison Project was launched in 1992 and is endorsed by the World Climate Research
Programme (WCRP) and the International Geosphere-Biosphere Programme (IGBP). Its
first phase focused on the Last Glacial Maximum (LGM) and the mid-Holocene (6 kyr BP)
and was designed to test and compare atmospheric components of climate models [Harrison
et al., 2002]. Phase I of PMIP met with considerable success, with 67 model experiments
archived in a central database and over 70 papers published in the refereed literature [Crucifix
et al., 2005]. One of the most important conclusions arising from this initial phase of PMIP
was the recognition of the importance of ocean and land surface (vegetation) feedbacks in
the climate system. In 2002, the PMIP steering committee decided to launch PMIP Phase
II which focuses on coupled ocean-atmosphere (OAGCM) and ocean-atmosphere-vegetation
models (OAVGCM). It also expanded the set of standard experiments to include simulations
of the previous interglacial/glacial transition (glacial inception at 115 kBP), the Younger
Dryas (12.7 - 11.7 kBP), the early Holocene (9 kBP) and the abrupt cooling event 8.2 kBP
[Crucifix et al., 2005].
Rapid Climate Change (RAPID), a six year program (2001-2007) of the Natural Environment Research Council aims to investigate the causes of rapid climate change, with a
main focus on the role of the thermohaline circulation. Using a combinations of present
day observations, paleoproxy data and climate models, RAPID seeks to assess the probability and magnitude of future climate change (Rapid Climate Science Plan, available at
http://www.noc.soton.ac.uk/rapid/rapid.php).
PAst Global changES (PAGES) is a core project of the International Geosphere-Biosphere
Program (IGBP) and is funded by the U.S. and Swiss National Science Foundations, and the
National Oceanic and Atmospheric Administration (NOAA). It acts as an umbrella organization to facilitate international collaborations and interdisciplinary science. PAGES does
9
not fund individual research, but it sponsors workshops, symposia and conferences. PAGES’
scope of interest includes the physical climate system, biogeochemical cycles, ecosystem processes, biodiversity, and human dimensions, on different time scales such as the Pleistocene,
the Holocene, the last millennium and the recent past. Over 3800 scientists in more than 100
countries around the world currently subscribe to PAGES (http://www.pages.unibe.ch/).
The Global Analysis, Interpretation and Modelling Task Force (GAIM) is a component
of the International Geosphere Biosphere Program (IGBP) of the International Council of
Scientific Unions (ICSU). GAIM aims to advance the study of the coupled dynamics of Earth
Systems with emphasis on the biogeochemical processes using both data and models. Modelling projects are structured by topic (CO2 , trace gases and climate-vegetation interactions)
and time period (paleo (from 20 kBP), fossil fuel (from 200 BP), contemporary (from 20
BP) and future). For example, the Ocean Carbon Cycle Model Intercomparison Project
(OCMIP) and the Coupled Carbon Cycle Climate Model Intercomparison Project (C4MIP)
are two of the GAIM continuing current projects (http://gaim.unh.edu).
The commission on Palaeoclimate (PALCOMM) is one of the five commissions of the
International Union for Quaternary Research (INQUA). Although mostly data oriented,
PALCOMM also assists the comparisons of paleo data and models. Its overall objective is a
better understanding of Quaternary climatic perturbations, transitions and abrupt events.
Finally, DEKLIM (German Climate Research Programme), a 37 million Euro project
funded by the Federal Ministry of Education and Research, dedicated one of its four major
research areas to Paleoclimate Research. The program seeks to (a) enhance our understanding of the climate system, (b) support young scientists and (c) invigorate international
cooperation.
5
Conclusion
Understanding the Earth’s past environments is not only a fascinating research field satisfying human curiosity, it is also essential in order to make predictions for the future. Today,
several international and interdisciplinary collaborations help to develop common international science directions and ensure that important scientific questions are addressed in a
coherent manner. Scientists in each of the two classic schools of paleoceanography (reconstructions through proxy data and theoretical quantitative analysis using models) are more
and more exchanging expertise and working together for a better understanding of the past
oceanographic environment. However, inherent problems associated with the poor spatial
and temporal resolution of proxy data as well as uncertainties related to the proxy data
themselves make their interpretation difficult and sometimes impossible. There is an urgent
need for the creation of new well-calibrated and time-resolved paleo data records. At the
same time, models are limited by resolved processes, resolution and the quality of boundary
conditions. Overall, there is much progress to be made in both fields. In spite of these problems, the geological data as well as model simulations provide a substantial set of results
which gives us some insight on how the ocean and the whole climate system functions. This
knowledge is of ultimate importance to understand and predict future climate changes due
to anthropogenic perturbations.
10
References
Bice KL, Scotese CR, Seidov D, Barron EJ (2000) Quantifying the role of geographic change
in Cenozoic ocean heat transport using uncoupled atmosphere and ocean models. Palaeogeography Palaeoclimatology Palaeoecology 161: 295–310
Bond G, Kromer B, Beer J, Muscheler R, Evans M, Showers W, Hoffmann S, Lotti-Bond R,
Hajdas I, Bonani G (2001) Persistent solar influence on North Atlantic surface circulation
during the Holocene. Science 294: 2130–2136
Broecker W, Lynch-Stieglitz J, Archer D, Hofmann M, Maier-Reimer E, Marchal O, Stocker
T, Gruber N (1999) How strong is the Harvardton-Bear constraint? Global Biogeochemical
Cycles 13(4): 817–820
Bryan F (1986) High-latitude salinity effects and interhemispheric thermohaline circulations.
Nature 323: 301–304
Budyko MI (1969) The effect of solar radiation variations on the climate of the earth. Tellus
21: 611–619
Bush AB, Philander SGH (1998) The role of ocean-atmosphere interactions in tropical cooling during the Last Glacial Maximum. Science 297: 1341–1344
Butzin M, Prange M, Lohmann G (2005) Radiocarbon simulations for the glacial ocean: The
effects of wind stress, Southern Ocean sea ice and Heinrich events. Earth and Planetary
Science Letters 235(1-2): 45 – 61
Clark PU, Pisias NG, Stocker TF, Weaver AJ (2002) The role of the thermohaline circulation
in abrupt climate change. Nature 415: 863–869
Clarke GKC, Lhomme N, Marshall SJ (2005) Tracer transport in the Greenland ice sheet:
three-dimensional isotopic stratigraphy. Quaternary Science Reviews 24: 155 – 171
Claussen M, Mysak LA, Weaver AJ, Crucifix M, Fichefet T, Loutre MF, Weber SL, Alcamo
J, Alexeev VA, Berger A, Calov R, Ganopolski A, Goosse H, Lohmann G, Lunkeit F,
Mokhov II, Petoukhov V, Stone P, Wang Z (2002) Earth system models of intermediate
complexity: closing the gap in the spectrum of climate models. Climate Dynamics 18: 579
– 586
CLIMAP Project Members (1976) The surface of the Ice-Age Earth. Science 191: 1131–1137
CLIMAP Project Members (1981) Seasonal reconstructions of the earth’s surface at the Last
Glacial Maximum. Geol Soc Am Map Chart Ser, MC-36
COHMAP Members (1988) Climatic changes of the last 18,000 years: Observations and
model simulations. Science 241: 1043–1052
Crowley TJ, North GR (1991) Paleoclimatology. 18, Oxford Monographs On Geology And
Geophysics, Oxford University Press, New York
11
Crucifix M (2005a) Distribution of carbon isotopes in the glacial ocean: a model study.
Paleoceanography 20: PA4020: 1–18
Crucifix M (2005b) Palaeoclimate modeling: achievements and perspectives. Predictability of
the Evolution and Variation of the Multi-scale Earth System, 21st Century Earth Science
COE Workshop, University of Tokyo
Crucifix M, Braconnot P, Harrison SP, Otto-Bliesner B (2005) Second phase of Paleoclimate
Modelling Intercomparison Project. EOS 86(28)
de Vernal A, Hillaire-Marcel C, Turon JL, Matthiessen J (2000) Reconstruction of seasurface temperature, salinity and sea ice cover in the northern North Atlantic during the
Last Glacial Maximum. Canadian Journal of Earth Sciences 37: 725–750
Duplessy JC, Labeyrie L, Juillet-Leclerc A, Maitre F, Duprat J, Sarnthein M (1991) Surface
salinity reconstruction of the North Atlantic Ocean during the Last Glacial Maximum.
Oceanol Acta 14: 311–324
Fichefet T, Hovine S, Duplessy JC (1994) A model study of the Atlantic thermohaline
circulation during the Last Glacial Maximum. Nature 372(6503): 252–255
Ganopolski A, Rahmstorf S (2001) Rapid changes of glacial climate simulated in a coupled
climate model. Nature 409: 153–158
Gates WL (1976a) Modeling the ice-age climate. Science 191: 1138–1144
Gates WL (1976b) The numerical simulation of ice-age climate with a global general circulation model. J Atmos Sci 33: 1844–1873
Gildor H, Tziperman E, Toggweiler RJ (2002) The sea-ice switch mechanism and glacialinterglacial CO2 variations. Global Biogeochemical Cycles 16(10.1029/2001GB001446)
Hansen J, Lacis A, Rind D, Russell G, Stone P, Fung I, Ruedy R, Lerner J (1984) Climate
sensitivity: Analysis of feedback mechanisms. Climate Processes and Climate Sensitivity,
vol. 29 of Geophy. Monogr., AGU, pp. 130–163
Harrison SP, Braconnot P, Joussaume S, Hewitt C, Stouffer RJ (2002) Fourth international
workshop of the Paleoclimate Intercomparison Project (PMIP): launching PMIP Phase
II. EOS
Heinze C, Maier-Reimer E, Winn K (1991) Glacial pCO2 reduction by the world ocean:
Experiments with the Hamburg carbon cycle model. Paleoceanography 6(4): 395–430
Imbrie J, Imbrie KP (1979) Ice Ages, Soving the Mystery. Enslow Publishers
Kaplan JO, Prentice IC, Knorr W, Valdes PJ (2002) Modeling the dynamics of terrestrial
carbon storage since the Last Glacial Maximum. Geophysical Research Letters 29(22): 31:
1–4
12
Knox F, McElroy MB (1984) Changes in atmospheric CO2 : Influence of the marine biota at
high latitude. J Geophys Res 89: 4629–4637
Kubatzki C, Claussen M (2003) Modelers and geologists join forces at workshop. EOS 84(9):
79
LeGrande AN, Schmidt GA, Shindell DT, Field CV, Miller RL, Koch DM, Faluvegi G, Hoffmann G (2006) Consistent simulations of multiple proxy responses to an abrupt climate
change event. Proc Natl Acad Sci 103(4): 837–842
Manabe S, Bryan K (1985) CO2-induced change in a coupled ocean-atmosphere model and
its paleoclimatic implications. J Geophys Res 90: 11689–11708
Manabe S, Stouffer RJ (1988) Two stable equilibria of a coupled ocean-atmosphere model.
J Climate 1(9): 841–866
McGuffie K, Henderson-Sellers A (2001) Forty years of numerical climate modelling. International Journal of Climatology 21: 1067–1109
Meissner KJ, Schmittner A, Weaver AJ, Adkins JF (2003) The ventilation of the North
Atlantic Ocean during the Last Glacial Maximum - a comparison between simulated and
observed radiocarbon ages. Paleoceanography 18(2): 1:1–13
Nairn AEM, Smithwick ME (1976) The Continental Permian in Central, West, and South
Europe, chap. Permian paleogeography and climatology. Boston: D. Reidel
Paillard D (1998) The timing of Pleistocene glaciations from a simple multi-state climate
model. Nature 391: 378–381
Parrish JT (1982) Upwelling and petroleum source beds, with reference to the Paleozoic.
American Association of Petroleum Geologists Bulletin 66: 750 –774
Parrish JT (1998) Interpreting Pre-Quaternary Climate from the Geologic Record. Columbia
University Press
Paul A, Schäfer-Neth C (2005) How to combine sparse proxy data and coupled climate
models. Quaternary Science Reviews 24: 1095 – 1107
Rahmstorf S (1995) Bifurcations of the Atlantic thermohaline circulation in response to
changes in the hydrological cycle. Nature 378: 145–149
Rahmstorf S (2002) Ocean circulation and climate during the past 120,000 years. Nature
419: 207–214
Rahmstorf S, Crucifix M, Ganopolski A, Goosse H, Kamenkovich IV, Knutti R, Lohmann
G, Marsh R, Mysak L, Wang Z, Weaver AJ (2005) Thermohaline circulation hysteresis: a
model intercomparison. Geophysical Research Letters 32, doi:10.1029/2005GL023655
13
Renssen H, Braconnot P, Tett SFB, von Storch H, Weber SL (2004) Past Climate Variability
through Europe and Africa, chap. Recent developments in Holocene climate modelling.
Kluwer Academic Publishers, Dordrecht, The Netherlands
Roche D, Paillard D, Ganopolski A, Hoffman G (2004) Oceanic oxygen-18 at the present
day and LGM: equilibrium simulations with a coupled climate model of intermediate
complexity. Earth and Planetary Science Letters 218: 317–330
Sarmiento JL, Toggweiler JR (1984) A new model for the role of the oceans in determining
atmospheric pCO2 . Nature 308: 621–624
Sarnthein M, Jansen E, Weinelt M, Arnold M, Duplessy JC, Erlenkeuser H, Flatoy A, Johannessen G, Johannessen T, Jung S, Koc N, Labeyrie L, Maslin M, Pflaumann U, Schulz
H (1995) Variations in Atlantic surface ocean paleoceanography, 50o - 80o N: A time-slice
record of the last 30,000 years. Paleoceanography 10(6): 1063–1094
Schmidt GA (1999) Forward modelling of carbonate proxy data from planktonic foraminifera
using oxygen isotope tracers in a global ocean model. Paleoceanography 14: 482–497
Schmittner A (2005) Decline of the marine ecosystem caused by a reduction in the Atlantic
overturning circulation. Nature 434: 628–633
Schmittner A, Meissner KJ, Eby M, Weaver AJ (2002) Forcing of the deep ocean circulation
in simulations of the Last Glacial Maximum. Paleoceanography 17(2): 5:1–16
Seidov D, Sarnthein M, Stattegger K, Prien R, Weinelt M (1996) North Atlantic ocean circulation during the Last Glacial Maximum and subsequent meltwater event: A numerical
model. J Geophys Res 101(C7): 16,305–16,332
Sellers WD (1969) A global climatic model based on the energy balance of the earthatmosphere system. J Appl Met 8: 392–400
Siegenthaler U, Wenk T (1984) Rapid atmospheric CO2 variations and ocean circulation.
Nature 308: 624–626
Stephens BB, Keeling RF (2000) The influence of Antarctic sea ice on glacial-interglacial
CO2 variations. Nature 404: 171–174
Stocker TF, Wright DG (1996) Rapid changes in ocean circulation and atmospheric radiocarbon. Paleoceanography 11(6): 773 – 795
Toggweiler JR (1999) Variation of atmospheric CO2 by ventilation of the ocean’s deepest
water. Paleoceanography 14(5): 571–588
Weaver A, Hughes TMC (1992) Stability and variability of the thermohaline circulation and
its link to climate. Trends in Physical Oceanography Research Trends Series Council of
Scientific Research Integration Trivandrum India 1: 15–70
Weaver AJ (1995) Driving the ocean conveyor. Nature 378: 135
14
Werner M, Mikolajewicz U, Hoffmann G, Heimann M (2000) Possible changes of δ 18 O in
precipitation caused by a meltwater event in the North Atlantic. J Geophys Res 105(D8):
10,161 – 10,168
Williams J, Barry RG, Washington WM (1974) Simulation of the atmospheric circulation using the NCAR global circulation model with ice age boundary conditions. J Appl Meteorol
13: 305 – 317
Winguth AME, Archer D, Maier-Reimer E, Mikolajewicz U (2002) Inverse Methods in Global
Biogeochemical Cycles, chap. Paleonutrient data analysis of the glacial Atlantic using an
adjoint ocean general circulation model. AGU Geophysical Mongraph Series, pp. 171–183
Wunsch C, ed. (1997) The ocean circulation inverse problem. Cambridge Univ. Press, Cambridge
Ziegler AM, Bambach RK, Parrish JT, Barrett SF, Gierlowski EH, Parker WC, Raymond
A, Sepkoski JJJ (1981) Paleobotany, Paleoecology, and Evolution Vol. 2, chap. Paleozoic
biogeography and climatology. New York: Praeger